1. Field of the Invention
The present invention relates to a system and method for shared directory management, and particularly to a system and method for shared directory failover management in a clustered server system.
2. Description of the Related Art
Clustered server systems are systems that integrate several independent servers. When a failure, such as system crash or memory damage occur in any one of the servers, other servers in the cluster take over the work executed on the failed server, so that the whole system can continue to provide service without interruption.
The shared directory is also called a public directory, namely the shared resources in general computers. When the directory on disk is set in a shared resource state, the directory becomes a shared directory, and the shared directory can be accessed by the other users via networks.
In the conventional method for shared directory management in clustered server systems, there are two ways to manage the shared directory. The first method sets each of the shared directories as a cluster resource (using one resource per share). The cluster resource represents a component that can be taken over by other servers if a failover (the action in which a failure occurs in one clustered server and the other clustered servers take over the work of the failed clustered server) occurs. In this method, since each of the shared directories is set as a cluster resource, the shared directories can be taken over when a failover occurs.
However, the number of resources controlled by the clustered server has a maximum limitation in practice. For example, Windows 2000 Advanced Server can control at most 1600 cluster resources in theory, and Windows NT 4.0 Server Enterprise Edition can control at most 800 cluster resources. Further, when more cluster resources are used by the system, the time for failover execution is longer.
The second method is to use only one cluster resource (using shared sub-directory options). A cluster resource can be set to indicate a specific directory, and each of the sub-directories in this specific directory will be shared. Therefore, there is no problem with maximum limitations of cluster resources.
However, in this method, the setting information of each of the sub-directories, such as commentary, access authority, maximum users and the like are identical to the specific directory. To change the setting information of each of the sub-directories, the system must provide other management applications for manual setting, thus increasing management costs. As well, the setting information of the sub-directories will be lost after being taken over if a failover occurs, further increasing the management complexity.